>-----Original Message----- >From: Jason Gunthorpe [mailto:jgg@xxxxxxxx] >Sent: Monday, November 20, 2017 3:26 PM >To: Shaikh, Azhar <azhar.shaikh@xxxxxxxxx> >Cc: jarkko.sakkinen@xxxxxxxxxxxxxxx; peterhuewe@xxxxxx; linux- >integrity@xxxxxxxxxxxxxxx >Subject: Re: [PATCH RFC v3 1/2] tpm: Keep CLKRUN enabled throughout the >duration of transmit_cmd() > >On Mon, Nov 20, 2017 at 09:19:37PM +0000, Shaikh, Azhar wrote: > >> -> tpm_tis_core_init() > >Yes, and IIRC, this covers tpm_tis_update_timeouts() too? > tpm_tis_update_timeouts() doesn't seem to be covered by tpm_tis_core_init(). I cannot find tpm_tis_update_timeouts() called from any function. >> -> tpm_tis_plat_remove() > >This should be in tpm_tis_remove and a little rework would be needed here > >> -> tpm_tis_reenable_interrupts() >> -> tpm_transmit_cmd() [ Already implemented in this patch ] > >Yes > >> >You could add a debug to ensure that read/write is never without >> >clk_enable being held. >> >> Sorry, but I didn't get this, adding a debug part. > >Just check a flag in the read/write functions to see if clkrun is on and WARN if >not > >Not sure if this is going to be overall better than what you have or not.. But >considering the nature of the bug it seems safer to have a wider range when >CLKRUN is working? > Having a wider range would mean clocks running for a bit longer duration than current implementation, which might have power implications, minimal though, would be good if we could avoid it from here :-) Also since this patch has already passed rigorous testing as mentioned earlier, hence would be good if we have the current implementation as it is. >Jason
